A new lawsuit adds more local and regional titles to the legal fight over how artificial intelligence was trained. On November 26, 2025, nine newspapers filed a copyright action in New York against OpenAI and Microsoft, alleging large scale copying of their work to build commercial AI tools. The filing names the Los Angeles Daily News, San Diego Union-Tribune, San Bernardino Sun, Boston Herald, Hartford Courant, The Morning Call, Boulder’s Daily Camera, the Daily Press, and The Virginian-Pilot. The case was opened in the Southern District of New York through a new complaint.
Damages sought exceed $10 billion
The plaintiffs outline a familiar arc, from alleged scraping and memorization of articles to AI outputs that replicate paywalled text. The publishers “seek damages in excess of $10 billion,” the complaint says. That sum equals roughly C$13.6 billion.
The filing also argues that product features built on large language models reduce referral traffic and subscriptions, weakening already thin local news economics. OpenAI and Microsoft are listed across multiple corporate entities, reflecting how models, APIs, and end user services are packaged and sold. Both firms have previously said they support publishers and are pursuing licences, while defending training as lawful fair use.
Prior case shapes the legal path
This action follows an earlier suit filed in April 2024 by other Alden Global Capital titles, including the Chicago Tribune and the New York Daily News. That case moved ahead in March 2025 when a federal judge allowed key claims to proceed, while trimming others.
“The claims the court has dismissed do not undermine the main thrust of our case,” said Frank Pine, executive editor for MediaNews Group and Tribune Publishing.
The new filing ties its facts to that consolidated docket, citing overlapping questions on training data, outputs, and attribution. It also highlights discovery fights already under way, which could influence how fast the new case moves.
For publishers, the stakes are direct revenue and long term bargaining power. If courts reject broad fair use arguments for training on copyrighted works without payment, licensing markets may expand quickly. If not, the path may hinge on product attribution and traffic sharing instead of cash licences.
In Canada, several major outlets, including CBC and Torstar, launched their own lawsuit in November 2024, signalling rising cross‑border pressure for negotiated deals. Public agencies and universities that procure AI tools are watching, since contract language on data sources and indemnities could shift with each ruling. The nine U.S. papers now add more data points to that landscape. A trial is not assured, but the damages ask, $10 billion (C$13.6 billion), raises settlement calculus for every party involved.